AI-driven radio interface under research of AGH University scientists

In the coming years, we will see the emergence of many new Internet applications, such as the meta-universe (augmented reality and virtual reality implementations), holographic telepresence, the Internet of Senses, the consolidation of the Internet of Things with autonomous robots, fully automated industries and manufacturing plants, and intelligent infrastructures and environments. To meet high communication requirements (in terms of throughput, latency, reliability, availability, and power consumption), wireless networks, and in particular their radio interface, are becoming extremely complex, with a myriad of advanced communication functions, protocols, and parameters that usually involve hidden dependencies between them. To deal with such complexity, the use of artificial intelligence and machine learning (AI/ML) techniques – and their ability to deal with complexity in general – is an essential factor in enhancing the efficiency of next-generation wireless networks.

The goal of the research team is to build a new AI/ML-driven radio (MLDR) interface, free from the limitations of existing wireless techniques. This new interface will learn to communicate by selecting and configuring a set of communication protocols and functionalities that fit each specific use case and scenario better, thus meeting the aforementioned performance requirements and making efficient use of available radio resources. In other words, MLDR will enable the deployment of "tailor-made" wireless networks.

The project titled “MLDR: A Machine Learning-Driven Radio Interface” received financing in the amount of EUR 600,000, of which PLN 871,080 will be granted to the research team from the AGH University. The research will be conducted by a consortium managed by a team from Spain (Pompeu Fabra University, Barcelona). The other consortium partners, besides the AGH University, are the University of Oulu (Finland) and CentraleSupelec (Paris, France).

***

The CHIST-ERA (European Coordinated Research on Long-term Challenges in Information and Communication Sciences & Technologies) network consists of research funding agencies. The aim of the network is to support research in ICST (Information and Communication Science and Technologies). To date, CHIST-ERA has announced 11 competitions for international research projects carried out jointly by teams from network member countries.

The CHIST-ERA Call 2022 competition was open to research projects conducted by international consortia involving at least three teams from at least three different countries participating in the competition. Joint proposals were evaluated by two international teams of experts, appointed separately for the two topics of the competition: Security and Privacy in Decentralised and Distributed Systems (SPiDDS) and Machine Learning-based Communication Systems, towards Wireless AI (WAI).

The CHIST-ERA network has selected a total of 12 international projects in the competition, with teams from 22 countries. Eight of the winning projects will be implemented within the SPiDDS topic, and four within the WAI topic. Three projects include teams from Poland represented by the AGH University, Poznan University of Technology, and the University of Gdańsk.